Dementia is a progressive impairment characterised by declining cognitive functions starting with minor memory problems but leading to deteriorating memory, orientation, language, and learning processes. In Aotearoa New Zealand, the estimated prevalence of dementia in 2020 was 1.4%, but in 2050, the number of people with dementia is expected to double. Since there is very little epidemiological information in New Zealand, routinely collected data could be a source of information to research the course of any disease, including dementia, in the country. We aim to develop machine learning models with routinely collected health data that enable the classification of dementia risk factors and the prediction of future dementia diagnoses based on ethical attitudes and preferences that people have about using their health data in dementia research. This study would give medical personnel another tool for early diagnosis of dementia, which would reduce costs and improve the quality of life of patients.
Keywords: Dementia, Longitudinal Data, Time Series